The new Nvidia speech to text model, Nemotron 3.5 Transcribe, is free and open-weight, covers up to 40 languages, and is built for recorded meetings and medical audio—not live streams.
Nvidia released Nemotron 3.5 Transcribe, a free open-weight model that can transcribe audio in up to 40 languages.
It can translate 31 languages to and from English, handle recordings up to an hour, and tell apart up to eight speakers.
The model works on recorded audio only, not live streams.
At about 30 billion parameters, Nvidia says it can run on a single chip, from Jetson AGX Thor and DGX Spark to Blackwell datacenter chips.
Nvidia also released version 2 of Granary, its open-source speech dataset, for developers customising apps.